Incorporating Wind Information in the Inversion of Co-Located Pressure and Seismic Data for Shallow Elastic Structure

摘要

When surface pressure is large, seismic noise between about 0.01 and 0.05 Hz is mostly generated by wind-related surface pressure changes. By analyzing this phenomenon using co-located pressure and seismic data, we can derive shallow elastic structure in the uppermost 50-100 m. However, there still remains a difficult question on seismic excitation by wind-related surface pressure because of complexity in wind behaviors. Roughly speaking, it consists of the mean shear flow and turbulence. The mean flow should act like a moving pressure source on the surface while turbulent parts act like a stochastic source as the surface pressure field becomes heterogeneous. We point out that this complexity may be reduced by selecting time intervals of stable wind direction. During such time intervals, pressure changes occur in the direction of winds, which justifies us to use a relatively simple, moving pressure source model. We quantify the stability of wind direction by the standard deviation of wind direction data (Delta psi). For stations with co-located wind, pressure, and seismic data, we propose to select time intervals with small Delta psi first and apply our inversion method based on the moving pressure model. In the absence of wind data at a station with co-located pressure and seismic data, we propose to analyze the high-pressure end of data, as this selection partially satisfies the stable wind condition.